Sun and co-authors measure for the first time, the nitrogen stable isotopic composition of ship emitted NO_x (δ¹⁵N-NO_x). They find a large range in values of ship emitted δ¹⁵N-NO_x that increased with tightening emission regulations and reduced atmospheric NO_x concentrations. Emission regulations are found to have the greatest influence on ship emitted δ¹⁵N-NO_x, which was explored using multiple statistical techniques. The biggest difference in ship emitted δ¹⁵N-NO_x occurring when IMO Tier III emission standards where implemented, and ships began adopting NO_x emission control technologies like selective catalytic reduction (SCR).

With transportation fast becoming one of the most important emission sources of anthropogenic NO_x, reliable values of ship emitted δ¹⁵N-NO_x are essential to accurate assessments of atmospheric NO_x sources, especially in coastal regions. A such, these data are valuable and should be considered for publication in ACP. While the data presented support the main conclusions (barring the erroneous Fig. 2), some additional information particularly regarding the implications of the data (Sect. 3.3) may benefit readers when interpreting the results, and those who intend to build upon this work for future studies. 

Major revisions stated below:

Figure 2 – Vessel category as opposed to emission regulation stages are shown on the x-axis. The means reported do not match the symbols (red squares). It appears the wrong categories are plotted here.  

Figure S9 – What do the yellow and dark blue sections of the 100% bar chart represent? Additionally It is unclear how the TSi was calculated based on the age distribution of ships during 2001 and 2021. Is there a maximum age beyond which ships no longer comply with Tier I, II or III emission regulations? Or is the age of the ship indicative of which emission standard it complies with? Clarification is needed.

Minor revisions stated below:

Some introductory text regarding the utilisation of stable isotope ratios is required, i.e., a description of delta notation and the units per mil (‰).

Line 36: Define NOx (NO + NO₂)

Line 37: Replace vital with important

Line 50: in comparison with the year 2000

Line 54 to 56: A concluding remark is required here. Something like, The current world merchant fleet thus comprises of more,  newly built vessels that have benefitted from the implementation of emission reduction technologies.  

Line 70: Primary contributor to what? Total anthropogenic NOx?

Line 71: Rather than state three major urban agglomerations, name the regions.

Line 74: Replace the phrase “in the face of”, with “due to”

Line 77: Is a powerful method used to apportion

Line 87: To address the lack of δ¹⁵N-NO_x measurements associated with ship emissions, rather than “Aiming at the knotty problem of lacking δ¹⁵N-NO_x”.

Line 110: Some description of how changes in meteorological conditions may affect measured δ¹⁵N-NO_x emitted from ships, or some references to previous work that has explored this would be beneficial so that the reader is aware of any potential implications. This should be done before noting that the effect is beyond the scope of the study, and will not be addressed further.

Line 116: but boilers were not sampled since the contribution of boiler exhaust to NOx emissions is weak.

Line 130: How was isotopic fractionation avoided?

Line 197: and NOx produced thermally

Line 212: suggesting

Line 223: Refer to Fig. 1 at the end of the sentence.

Line 224 to 226: I found this concluding sentence slightly confusing a little vague. Do you mean that given ship emitted δ¹⁵N-NO_x differs substantially to the δ¹⁵N-NO_x produced by diesel, gasoline and LPG powered combustion engines, further investigation is required to determine the factors influencing ship emitted δ¹⁵N-NO_x  for the accuracy of source apportionment etc. If so, please describe as such, alternatively please clarify the concluding sentence here.

Figure 1: For figure one, there is no need to include the word “values” in the y-axis label. The colors of each box and whisker diagram could cause confusion, as I originally assumed like colors were related in some way (i.e., similar types). If you are not able to use a different colors for each type, please consider having all ship types (Ship, Ship 6.1 and Ship 8.7) in the same color, all gasoline type (gasoline and gasoline 5.2) etc. in the same colors. It would also be useful to indicate on the face of the figure, which types have been adjusted for not having SCR, perhaps with square brackets above or beneath the relevant box and whisker diagrams.

Line 246: These classification indicators are considered because, as opposed to “The consideration is because”.

Line 254: Instead of (small p values), (indicated by p values < xxx) would be more descriptive.

Line 255: By “divided by” do you mean between the two indicators?

Line 254 to 256: This sentence is slightly confusing, do you mean: Similar significant differences (small p values) in δ¹⁵N-NO_x values between types of emission regulations, and different vessel categories where calculated by the Mann-Whitney U test, as shown in Fig. 2 and Fig. S2, respectively. If so, please clarify.

Line 283 to 285: It is unclear to me what great discrepancy is being referred to here. Please clarify.

Line 287: Instead of “The insignificant discrepancy of the δ¹⁵N-NO_x values from ships implementing Tier I and Tier II”, “The insignificant difference between δ¹⁵N-NO_x values from ships implementing Tier I and Tier II” is clearer.

Line 304: Please refer to Fig. 1 at the of the sentence here.

Line 319: Please clarify “The discrepancy”, i.e., “The stronger relationship between δ¹⁵N-NO_x and NOx concentration for vehicles with NOx emission control technologies versus vehicles without, is attributed to the enrichment of δ¹⁵N..”

Section 3.2 would benefit from a table summarising the 4 emissions regulation stages, whether Tier I, II or III was implemented, what the major differences are (e.g., SCR/ fuel optimisation and precombustion control technologies), the corresponding δ¹⁵N-NO_x and NOx concentrations, and emission factors of NOx pertaining to section 3.3. This will also be helpful to section 3.3 to summarise the parameters set for the mass weighted model (i.e, emission factors and δ¹⁵N-NO_x values).

Line 381: Please include a description of units either here, or in the table.

Review of Sun et al.

This work reports on measurements of the isotope composition of NOx emitted from ships under a variety of operating mechanisms, fuel types, ship types/ages, etc. The results are reported both in terms of the direct observations as well as in the context of weighting by emissions and correcting for fractionations associated with scrubbing technology. The study also predicts the changes in the isotope composition of ship emissions over time, which is very useful for future studies. It is a very well-done study and overall the results are compelling. The comments overall are minor, but important to the scientific merit of the work.

The manuscript appears written for a short form journal and would benefit from a fuller discussion of implications and a conclusions section. There also several points recommended below to pull in text and/or figures from the supplement since there is space to do so for ACP. Most of the comments surround clarification of weighting specifics for calculating the final isotopic composition, making recommendations on how best to use the reported values for future studies and clarifying meaning of text in several places.

Title: I appreciate the “first measurements” specificity here, but I think the authors do more than just measure the d15N. I would suggest using the word “characterization” and also making it clear that these values apply to emissions. For example: “Characterization of the nitrogen stable isotope composition (d15N) of ship NOx emissions”

Abstract:

The abstract is a good representation of the study. It would be helpful to clarify the values presented. Are these the directly observed values or weighted values, specify this. At the end of the abstract, it would also be good to acknowledge that you present a framework to compute accurate assessments over time.

Introduction:

In the introduction, it would be useful to give more background information about how much NOx is emitted under different operating conditions (i.e. why are three operating conditions specifically chosen, how are they regulated, etc). As well as a clearer description or table of the emissions standards – i.e. Tier I, Tier II, Tier III. I was not familiar with much of the policy and regulatory framing before reading this paper. Additionally, it would be helpful on line 50 to define MARPOL and give a bit more information in a table about the emissions regulations and on which dates these changed (the predicted time series in Figure 3 is based on this).

Methodology:

It would be helpful to have a photo or diagram of the set up on the ship (this could go in the supplement). It is difficult to picture the setup based on the limited description here.

On line 112 it is stated that any impact of meteorological conditions is ignored. I suspect this fine given that the sampling is taking place within the smokestack. However, could the authors provide evidence for a lack of systematic differences for summer vs winter conditions for instance? This will come up again later, but there is little discussion of the internal variability of the signals within an operating condition (e.g. operating under cruising mode produces a large range of values for the same ship emissions) and it is curious whether the emissions change at all within a specific operating condition due to other factors such as meteorology or the respond of the engine to outside temperature conditions.

On line 145-146 background and blank samples are collected but the data from these samples is not discussed in the results and should be. On Line 160 it is stated that the average blank concentration was ~1.15 % -- is this a lab blank? A field blank? Or an N2O blank associated with the denitrifier method? How much average change results from correcting for the blank overall?

Equation 2: How was this equation determined? Is it determined empirically? (ie based on data in this study). What is LF to the power of 3? And what units are used for each variable in this equation?

Results and discussion:

The reported range of measurement is -35.8 per mil to 2.0 per mil. Are these the initial data or have these already been weighted based on Eq. 2? In order to re-calculate these values, the AE and ME fractions also need to be reported. Can this be included in a table in the supplement?

Suggest bringing Figure S7 into the main manuscript in Section 3.1. Also, consider formatting the symbols to be colored by whether the ship was operating under Tier I, Tier II or Tier III emissions regulations. It would be helpful to get a sense of the variability in concentration with the samples and this is not currently reported in the main manuscript.

This section should also include more discussion of the variability within an observed group – ie why is there such large variation within the same ship emissions under hoteling condition?

Is the distance from shore important at all for variation in d15N-NOx?  

Line 224: This last sentence does not follow from the evidence presented. The comparison suggests that it is necessary to know the types of engine in order to accurately apportion sources or to treat the mix of engines with respect to the fact that they each have different values associated with their NOx emissions.

Furthermore, be sure it is appropriate to directly compares values here. Some methods may only collect NO2 while the methodology used here collects both NO and NO2.  

Starting page 10, the discussion of Figure S4 and the conclusions drawn from this analysis are a bit unclear/difficult to follow. More of the discussion and interpretation from Text S1 should be included here. Also consider including Figure S4. And parts of Fig S4 are cutoff and unreadable, which also adds to the confusion. (Also note that on line 276, the stages are called I, II, III etc instead of 1, 2, 3… this may confused readers bc of the Tier # system. Also Fig S4 does not show Stage 3 – why not?).

Figure 2/Line 305: the x-axis does not really plot emission regulation. It’s really ship category correct? I understand that different regulations apply to different types of ships, but the emission regulations are not actually stated on this plot. Also, what does the large x-range mean on the plots here? For instance, the passenger ship shows a significant bellowing out in the horizontal direction (as does the research ship results), but there is no quantity on the x-axis, so what does this range signify here?

Section 3.3: This study should put forward recommendations on what values should be used in current studies for ship emissions. How sensitive is the d15N from ship emissions to regional differences in ship fleets? Are there regional different in ship fleets? For instance, perhaps it  could be calculated what regional values are so coastal studies in different locations could use those values. Or make a recommendation on how to calculate a value for your specific study region/study conditions.

The manuscript should also include a final conclusions section.

Minor technical comments:

Line 57-58: The last sentence doesn’t follow and then is repeated in the next paragraph. Suggesting removing sentence starting with “The rapid development…”

Line 72: replace ‘continuous updates’, i.e., “These studies of ship emissions and their environmental impacts…”

Line 74: replace ‘some other’ with ‘additional’

Line 77: please define delta units here d15N does simply equal the ratio of 15N/14N

Line 81: change ‘emitted from’ to ‘characterized for’

Line 98: change ‘researches’ to ‘research’

Line 114: suggest changing ‘emitted d15N-NOx’ values to ‘emitted NOx’.

Line 117: this sentence is awkward. Suggest rephrasing to: “NOx emitted from AE were also collected, but the boiler were not sampled since these make a small contribution to NOx emissions…”

Line 149: Since for all sample redundant KMnO4 was removed this line appears awkward. Please rephrase.

Line 153: what is an “absorption solution”? I think it meant the sample that was collected as nitrate? Replace this phrase.

Line 198: add ‘NOx’ so it reads – “…and NOx produced thermally by internal combustion engines…”. Also should probably add the Snape paper that is cited in Walter’s et al here or just say “and references therein”.  

Line 247: can it be mentioned what some “other indicators” are that have been used (as examples)?

Line 279: What are GB I to GB VI? These are not previously defined.

Line 282-283: What is the “great discrepancy”? What does “hard to analyze” mean? This sentence over should be rephrased for clarity.

Line 321: please make this more clear when stating the strength of correlation is “just within range” – does that mean the r values are similar?

Line 324: ulteriorly is not correct here – perhaps “clearly”?

Line 326: is “thermodynamic” here the same as “thermal” which was previously used in the text?

Line 339: suggest changing to: “The catalytic reduction system is a major reason for the differences…”

Line 383: What is Imo, 2019?

Summary:  The authors present an interesting study on characterizing the nitrogen stable isotope composition (δ¹⁵N) of nitrogen oxides (NO_x) from ship emissions.  This work represents an important contribution to the field because ship emissions are an important source of NO_x; however, the δ¹⁵N values of this emission source had been previously unknown.  The authors have characterized the δ¹⁵N-NO_x from ship emissions from several different types of ship types that are also categorized by different emission standards.  Overall, the authors find that the δ¹⁵N-NO_x from ship emissions is extremely variable, similar to direct tailpipe emissions from light-duty vehicles.  There is a strong dependency on the δ¹⁵N-NO_x from ships that depend largely on the emission regulation standard and categorization of the ships.  The authors then model the expected change in δ¹⁵N-NO_x from ships emission as a function of emission regulations, helping to constrain the δ¹⁵N-NO_x emission signature.  This manuscript presents novel measurements and is generally suitable for publication in ACP.  My largest gripe with this work is that a lot of useful material and figures are reported in the Supplement, which is critical to understanding the work, and parts of the Supplement should be moved into the main text.  My recommendation is publication after minor revisions.  My specific comments are provided below.

Comments:

Line 21-23:  You may also mention that δ¹⁵N-NO_x from ship emissions could also be important for source apportionment of atmospheric nitrogen deposition in remote ocean regions.

Lines 31-34: Besides NO_x, it could also be useful for tracking its influence in atmospheric nitrate air quality and N deposition studies.

Lines 104-105:  Table S1 could be useful to include in the main text.  I was left wondering about the emission regulation categorization of the various types of ships throughout the manuscript, and having a table to reference this information would be extremely useful.

Lines 110-112: Whether meteorological conditions would impact δ15N-NOx is unclear.  Can the authors expand on this point or delete this sentence entirely?

Lines 115-117:  Can the authors quantify the relative contribution of the boiler emissions to the main and auxiliary engines?  Is it, for example, less than 5 %?

Lines 117-119:  Reactive gases can collect on stainless steel.  How do the authors expect the sampling apparatus to impact the isotope results?  Also, was water formed/condensed on the sampling apparatus?  Was there a relationship between collection time and the measured δ¹⁵N-NO_x?

Lines 126-128:  Can the collection time for each sample be provided in the Supplement?

Lines 149-151:  What was the precision of the NO₃^- concentration measurement?

Lines 153-155:  You may mention that this particular strain of bacteria lacks the N₂O reductase enzyme, which is critical for the analysis.

Lines 160-162:  Can you provide the average blank concentration and the number of blank samples taken in addition to the blank fractional contribution?  Additionally, what was the measured δ¹⁵N of the blank?

Line 174:  This equation should be modified to more clearly indicate which molecule δ¹⁵N corresponds to.

Lines 178-187:  These statistical tests are not common, and a description should be provided in the text, not the Supplement.  

Lines 190-191:  It would be nice to reference a figure when discussing/presenting the data.

Line 199-  You should expand on the kinetic isotope effect mentioned in this line to the kinetic isotope effect associated with the thermal decomposition of the strong triple bond of N₂.

Lines 251-252:  The outcomes of the variance tests are interesting and dictate a lot of the discussion in the upcoming section.  Thus, I would recommend including Table S5 in the main text.

Figure 272-273:  Figure 2 is mentioned to present the δ¹⁵N-NO_x from ships under different emission regulations, but instead, Figure 2 shows the categorization of ship type.  The δ¹⁵N-NO_x by emission regulations is shown in Figure S2.  Since these two factors are suggested to be the dominant driver of δ¹⁵N-NO_x and are thoroughly discussed in this section, I recommend combining Figure 2 and S2 into one figure showing the δ¹⁵N-NO_x relationship with both parameters.

Lines 329-334:  This enrichment factor, however, depends on how much NOx is reduced, and a shift inδ¹⁵N, which appears to have been applied to the measured data set, may not be appropriate.

Lines 359-402:  I found this section interesting but was left with a few questions that I hope the authors can comment on – will there be large spatial variability in the predicted δ¹⁵N-NO_x?  Also, how might the constrained δ¹⁵N-NO_x ship values impact previous source appointment studies?

  

Technical Comment:

TC1:  Isotope deltas (δ) are a quantity symbol and should be italicized throughout the manuscript.  

TC2:  The “x” in NO_x should be italicized throughout the manuscript.  

TC3:  All units, including ‰, should be separate from the value by a space.